drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris Westin (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (DRILL-2626) org.apache.drill.common.StackTrace seems to have duplicate code; should we re-use Throwable's code?
Date Mon, 30 Mar 2015 19:10:56 GMT

     [ https://issues.apache.org/jira/browse/DRILL-2626?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Chris Westin updated DRILL-2626:
    Affects Version/s: 0.8.0

> org.apache.drill.common.StackTrace seems to have duplicate code; should we re-use Throwable's
> ---------------------------------------------------------------------------------------------------
>                 Key: DRILL-2626
>                 URL: https://issues.apache.org/jira/browse/DRILL-2626
>             Project: Apache Drill
>          Issue Type: Bug
>    Affects Versions: 0.8.0
>            Reporter: Daniel Barclay (Drill)
>            Assignee: Chris Westin
> It seems that class org.apache.drill.common.StackTrace needlessly duplicates code that's
already in the JDK.
> In particular, it has code to format the stack trace.  That seems at least mostly redundant
with the formatting code already in java.lang.Throwable.
> StackTrace does have a comment about eliminating the StackTrace constructor from the
stack trace.  However, StackTrace does _not_ actuallly eliminate its contructor from the stack
trace (e.g., its stack traces start with "org.apache.drill.common.StackTrace.<init>:...").
> Should StackTrace be implemented by simply subclassing Throwable?  
> That would eliminate StackTrace's current formatting code (which would also eliminate
the difference between StackTrace's format and the standard format).
> That should also eliminate having the StackTrace constructor's stack frame show up in
the stack trace.  (Throwable's constructor/fillInStackTrace already handles that.)
> (Having "StackTrace extends Throwable" isn't ideal, since StackTrace is not intended
to be a kind of exception, but that would probably be better than the current form, given
the bugs StackTrace has/has had (DRILL-2624, DRILL-2625).
> That non-ideal subclassing could be eliminated by having a member variable of type Throwable
that is constructed during StackTrace's construction, although that would either cause the
StackTrace constructor to re-appear in the stack trace or require a non-trivial workaround
to re-eliminate it.
> Perhaps client code should simply use "new Throwable()" to capture the stack trace and
a static methods on a utility class to format the stack trace into a String.)

This message was sent by Atlassian JIRA

View raw message